Generating PD data from electrical treeing in silicone rubber for insulation lifetime modelling

Binti Ab Aziz, Nur Hakimah and Judd, Martin and Catterson, Victoria (2013) Generating PD data from electrical treeing in silicone rubber for insulation lifetime modelling. In: 7th Universities High Voltage Network Colloquium, 2014-01-15 - 2014-01-16.

Full text not available in this repository.Request a copy


Electrical treeing is a degradation phenomenon in solid dielectric material resulting from high and non-uniform electrical field or partial discharges (PD). The presence of electrical treeing, therefore, can be examined through PD monitoring by looking for characteristic features within the phase-resolved plot of PD data. As electrical trees evolve in time, time-resolved analysis of PD data may be more descriptive of the correspondence between discharges and tree propagation. Continuing partial discharges in electrical treeing may lead to catastrophic failure, but there is still a lack of understanding of the evolution of PD characteristics prior to breakdown. This paper focuses on a method of simplifying the production and growth of electrical trees in silicone rubber (SiR), an advanced insulating material that is widely used in high voltage cable accessories due to its excellent insulation and mechanical performance. Crucially, commercially available pre-formed samples of SiR are used to ensure consistency and eliminate the need for the mixing, degassing and heating process in sample preparation. The experimental methodology is described, in terms of sample preparation, applied voltage regime, and data capture. A constant 50 Hz AC voltage is applied to the samples (with a needle-plane test arrangement using hypodermic needles) at a level sufficient to induce PD, leading to breakdown within hours. Both IEC 60270 electrical method and radio frequency (RF) sensors are used to capture PD data, while a digital microscope is used for visual observation. The paper describes the features found within the PD phase-resolved plot, and evaluates the similarities and differences between the two measurement techniques. Future work aims to automatically detect those features corresponding to electrical tree growth, and give a lifetime prediction for the insulation samples being studied.